心音的分析

The Analysis on Heart Sounds

自上世纪70年代以来,由于计算机和现代数字信号处理技术的发展,生物医学工作者对心音进行了大量的分析研究熏相继运用了谱分析、时频分析、小波变换和最优匹配法等方法。谱分析不适合对非平稳随机心音信号的分析。国内外工作者通过加核函数和采用模糊函数等方法来减少时频分布中的交叉项。具有自适应性时—频窗的小波变换可以得到更能反映心音病例特征的信息。最优匹配法是一种没有交叉项的时频分析方法。

In this paper, the situation and the destination of the heart sound study were introduced, at the same time, some methods which were used to research heart sound by these biomedical researchers since the 70's on the last century were introduced, they included the spectral analysis, the distribution of time-frequency, the wavelet translation, the matching pursuit, and so on. The spectral analysis is not suit for analyzing the typical non-stationary stochastic signal. These researchers on the world have reduced the cross-terms on the distribution of time-frequency by nuclear function or fuzzy function. The wavelet translation with the adaptive time-frequency window could gain the characters that more reflect the symptoms of heart sound. The matching pursuit is the analysis of time-frequency without the cross-terns.